原创 How to Be a Good Graduate Student(2)

Doing Research

For   many   new   graduate   students,   graduate   school   is   unlike   anything   else   they've   done.   Sometimes   it's   hard   to
know exactly what it is you're supposed to be learning. Yes, you have to complete a dissertation, but how do you
start? What should you spend your time doing?

----------------------- Page 3-----------------------
Graduate school is a very unstructured environment in most cases. Graduate students typically take nine hours or
less of coursework per semester, especially after the second year. For many, the third year -- after coursework is
largely   finished   and   preliminary   exams   have   been   completed   --   is   a   very   difficult   and   stressful   period.   This   is
when you're supposed to find a thesis topic, if you're not one of the lucky few who has already found one. Once
you do find a topic, you can expect two or more years until completion, with very few landmarks or milestones in
sight.

The following sections talk about the day-to-day process of doing research, criticism and feedback, working on
the thesis, and financial support for research.

The Daily Grind

Being   a   good   researcher   involves   more   than   ``merely''   coming   up   with   brilliant   ideas   and   implementing   them.
Most researchers spend the majority of their time reading papers, discussing ideas with colleagues, writing and
revising papers, staring blankly into space -- and, of course, having brilliant ideas and implementing them.

A later section discusses the process and importance of becoming part of a larger research community, which is a
critical aspect of being a successful researcher. This section contains ideas on keeping track of where you're going,
and where you've been, with your research, staying motivated, and how to spend your time wisely.

Keeping   a   journal   of   your   research   activities   and   ideas   is   very   useful.   Write   down   speculations,   interesting
problems, possible solutions, random ideas, references to look up, notes on papers you've read, outlines of papers
to write, and interesting quotes. Read back through it periodically. You'll notice that the bits of random thoughts
start   to   come   together   and   form   a   pattern,   often   turning  into   a   research   project   or   even   a   thesis   topic.   I   was
surprised, looking back through my journal as I was finishing up my thesis, how early and often similar ideas had
cropped up in my thinking, and how they gradually evolved into a dissertation.

You'll have to read a lot of technical papers to become familiar with any field, and to stay current once you've
caught up. You may find yourself spending over half of your time reading, especially at the beginning. This is
normal.   It's   also   normal   to   be   overwhelmed   by   the   amount   of   reading   you   think   you   ``should''   do.   Try   to
remember that it's impossible to read everything that might be relevant: instead, read selectively. When you first
start reading up on a new field, ask your advisor or a fellow student what the most useful journals and conference
proceedings are in your field, and ask for a list of seminal or ``classic'' papers that you should definitely read. For
AI researchers, a useful (if slightly outdated) starting point is Agre's (see [agre]) summary of basic AI references.
Similar documents may exist for other research areas -- ask around. Start with these papers and the last few years
of journals and proceedings.

Before   bothering   to   read   *any*   paper,   make   sure   it's   worth   it.   Scan   the   title,   then   the   abstract,   then   --   if   you
haven't completely lost interest already -- glance at the introduction and conclusions. (Of course, if your advisor
tells you that this is an important paper, skip this preliminary step and jump right in!) Before you try to get all of
the nitty-gritty details of the paper, skim the whole thing, and try to get a feel for the most important points. If it
still seems worthwhile and relevant, go back and read the whole thing. Many people find it useful to take notes
while they read. Even if you don't go back later and reread them, it helps to focus your attention and forces you to
summarize as you read. And if you do need to refresh your memory later, rereading your notes is much easier and
faster than reading the whole paper.

A few other points to keep in mind as you read and evaluate papers:

       1.Make sure the ideas described really worked (as opposed to just being theoretically valid, or tested on a
          few toy examples).

      2.Try to get past buzzwords: they may sound good, but not mean much. Is there substance and an interesting
          idea underneath the jargon?

      3.To really understand a paper, you have to understand the motivations for the problem posed, the choices
          made in finding a solution, the assumptions behind the solution, whether the assumptions are realistic and
          whether they can be removed without invalidating the approach, future directions for research, what was
          actually   accomplished   or   implemented,   the   validity   (or   lack   thereof)   of   the   theoretical   justifications   or

----------------------- Page 4-----------------------
          empirical demonstrations, and the potential for extending and scaling the algorithm up.

Keep the papers you read filed away so you can find them again later, and set up an online bibliography (BibTeX
is a popular format, but anything consistent will do). I find it useful to add extra fields for keywords, the location
of   the   paper   (if   you   borrowed   the   reference   from  the   library   or   a   friend),   and   a   short   summary   of   particularly
interesting   papers.   This   bibliography   will   be   useful   for   later   reference,   for   writing   your   dissertation,   and   for
sharing with other graduate students (and eventually, perhaps, advisees).